Google News Monitor: Keyword & Topic Alerts via RSS

Track any keyword, company, person or topic on Google News and get clean, structured JSON: headline, publisher, publish time, Google News link, optional direct publisher URL, related coverage and more. The Actor reads the public Google News RSS feeds (no sign-in) in 71 countries and 40 languages, supports Google News search operators ("exact phrase", OR, -exclude, site:, intitle:), time ranges from the past hour to custom dates, top stories, topic sections and local news.

Turn on Only new articles and schedule the Actor: every run returns just the articles you haven't received yet. That's a news alert feed for Slack, email, a spreadsheet or a database, and you pay only for new articles.

Price: $1.50 per 1,000 articles ($0.0015 per article) plus $0.0005 per run. No subscription, no extra usage fees. Platform and proxy costs are included. Error rows and the run summary are free.

Unofficial. This Actor is not affiliated with, endorsed by or sponsored by Google. Google News is a trademark of Google LLC. See Legal before you use the data.

What you get

• One item per article with the same fields for every feed: title, publishedAt (ISO 8601 UTC), sourceName, sourceDomain, sourceUrl, googleNewsUrl, publisherUrl (optional), snippet, relatedArticles, articleId, language, country, query, position, isNew, scrapedAt, plus status (ok for an article, error for a free error row) and error / errorKind on error rows.
• Search queries with Google News operators, topic sections (Top stories, World, Nation, Business, Technology, Entertainment, Sports, Science, Health), local news for places, or any news.google.com URL (for example a company topic page).
• Time filters: past hour, 24 hours, 7 days, 30 days, or a custom date range. Every article's publish date is checked again, so nothing older slips through.
• More than 100 articles per query: Google caps each feed at about 100 articles; "Split long time ranges by day" fetches one feed per day (tested: 711 articles for one query over 7 days).
• Alert mode (onlyNew) that remembers what each query already returned, across runs.
• De-duplication across queries in a run (same article ID, or same headline from the same publisher).
• Publisher URLs (optional, best effort): decode Google News links into the publisher's own article URL.
• Clear errors: a query, place or URL that can't be read gets a free error row with the reason; the other queries still run.
• Built for automation: a predictable schema, pay-per-event pricing without usage fees, limited permissions, a SUMMARY record, and a clean stop when your spending limit or the run's time limit is reached.

Use cases

• Brand and PR monitoring: follow mentions of your company, products and executives in the countries and languages you care about. Schedule hourly with "Only new articles" for near-real-time alerts.
• Competitor news: watch competitors' launches, funding, hires and lawsuits with queries like "Acme Corp" OR "Acme Inc" -jobs.
• Finance and market news: track companies, sectors, central banks or commodities, e.g. "interest rates" site:reuters.com or intitle:Nvidia.
• Daily digests and newsletters: collect the day's headlines with publisher, date and link into a spreadsheet, an email or a team channel.
• Media research and dashboards: count coverage by publisher, country or day; compare editions (e.g. US vs. India vs. Germany).
• Local news: monitor a city or region with the local-news feed.

How to use

1. Add one or more Search queries (one per line), and/or pick Topics, Local news places or Google News URLs.
2. Choose the Country and Language (default: United States, English).
3. Pick a Time range (default: past 7 days) or a custom From / To date.
4. Set Max articles per query (default 100; 0 = everything Google returns).
5. For alerts: turn on Only new articles, save the input as a task and create a Schedule for it in Apify Console (e.g. every hour). Connect the results to Slack, email, Google Sheets, Zapier, Make or a webhook with Apify integrations.
6. Download the results as JSON, CSV, Excel or via the API. A run summary is saved as the SUMMARY record in the run's key-value store.

Google News search operators (verified)

Operator	Example	Meaning
"..."	"electric vehicles"	Exact phrase
OR	Nvidia OR AMD	Either term
-	Tesla -stock	Exclude a word
site:	tariffs site:reuters.com	Only one publisher's site
intitle:	intitle:Nvidia	Word must be in the headline
when:	apple when:1h	Time range (the Time range field adds this for you)
after: / before:	apple after:2026-10-01 before:2026-10-05	Date range (the From / To date fields add these for you)

If your query (or a pasted Google News URL) already contains when:, after: or before:, your operator is used for it instead of the Time range / date settings.

A google.com News tab link (https://www.google.com/search?q=...&tbm=nws) pasted into Google News URLs is turned into a search query.

Input example

{
  "queries": ["\"Acme Corp\" OR \"Acme Inc\"", "Tesla -stock", "\"interest rates\" site:reuters.com"],
  "topics": ["TOP_STORIES", "BUSINESS"],
  "locations": ["London"],
  "country": "US",
  "language": "auto",
  "timeRange": "1d",
  "maxArticlesPerQuery": 0,
  "onlyNew": true,
  "trackingId": "my-news-alerts",
  "deduplicate": true,
  "resolvePublisherUrls": false
}

Output example

A real item from a cloud test run of this version (query Federal Reserve, with Resolve publisher URLs on; long IDs shortened):

{
  "query": "Federal Reserve",
  "feedType": "search",
  "status": "ok",
  "position": 2,
  "title": "The Fed's Beige Book is Mentioning Diesel a Lot More Than Before. What That Could Mean for Consumer Prices.",
  "publishedAt": "2026-10-19T17:57:10Z",
  "sourceName": "International Business Times",
  "sourceDomain": "ibtimes.com",
  "sourceUrl": "https://www.ibtimes.com",
  "googleNewsUrl": "https://news.google.com/rss/articles/CBMiugFBVV95cUxPTlNpTEx4d2hG...?oc=5",
  "publisherUrl": "https://www.ibtimes.com/feds-beige-book-mentioning-diesel-lot-more-before-what-that-could-mean-consumer-prices-3808735",
  "snippet": "The Fed's Beige Book is Mentioning Diesel a Lot More Than Before. What That Could Mean for Consumer Prices. - International Business Times",
  "relatedArticles": [],
  "articleId": "CBMiugFBVV95cUxPTlNpTEx4d2hG...",
  "language": "en",
  "country": "US",
  "isNew": null,
  "scrapedAt": "2026-10-19T23:58:03Z",
  "error": null,
  "errorKind": null
}

Items from Top stories and topic sections usually also list relatedArticles: other publishers' coverage of the same story, each with title, sourceName and googleNewsUrl.

Output fields

Field	Description
query, feedType	The query, topic code (e.g. BUSINESS), place or URL that found the article, and search / topic / location / url
status	ok for an article (charged), error for a free error row
position	1-based position within this query's results in this run (after sorting)
title	Headline, without the " - Publisher" suffix Google adds
publishedAt	Publish time from the feed, ISO 8601 UTC
sourceName, sourceDomain, sourceUrl	Publisher name, domain and home page (from the feed's <source> element)
googleNewsUrl	Link on news.google.com; opens the article in a browser
publisherUrl	Direct article URL on the publisher's site, when Resolve publisher URLs is on and decoding worked; otherwise null
snippet	Plain text of the feed's description (see FAQ: Google no longer puts article summaries in its feeds)
relatedArticles	Other coverage of the same story (topic and top-story feeds)
articleId	Stable Google News article ID (same ID on every run)
language, country	The Google News edition the article came from
isNew	true in alert mode (every returned article is new to this alert memory), otherwise null
scrapedAt	When this run read the feed
error, errorKind	Error rows only: why the feed returned nothing, and invalid_input, not_found, blocked, source_error or internal_error

The dataset has three views: Articles (overview table), All fields and Errors.

Error rows (free)

If a query, topic, place or URL can't be read, the Actor adds one free row for it (status: "error") with the reason, and the other entries still run. Examples: an unknown place ("Google News has no feed for this topic, place or URL"), a URL that isn't on news.google.com, a link to a single article instead of a feed, or a feed Google blocked. An invalid setting for the whole run (for example an unknown country, a bad date or an unsupported proxy group) gives one error row with feedType: "input". Error rows are never charged.

Alert mode: only new articles

With Only new articles on, the Actor stores a short hash of every article it returned for each query, topic, place or URL in a named key-value store in your Apify account (google-news-monitor-seen). It stores only hashes, never article data. Each later run skips those articles.

Real test (8 feeds, past 24 hours, US edition):

Run	Articles returned	Already seen (skipped, free)
1st run	467	0
2nd run, 1 minute later	0	483
3rd run, 7 minutes later	42	448

Good to know:

• "New" means not returned to you before, not "published since the last run". An older article that newly enters Google's top results for your query is returned once.
• Articles are only remembered once they are actually delivered. If a run stops at your spending limit, the rest are returned by the next run.
• With Remove duplicates across queries on, an article returned for one query is not returned again for another query, in this run or later runs.
• Tracking ID: leave it empty and each saved task gets its own memory (a run without a task gets one memory per set of queries plus country and language). Set your own ID to share one memory between tasks, or to keep the memory when you change the query list. Monitors that ran before version 1.1 without an ID keep what they already received (the old shared memory is read too).
• Changing the time range or max articles doesn't reset the memory. Use a new Tracking ID to start fresh.
• An article is remembered until it has been missing from its feed for 180 days (up to 20,000 articles per query). Articles that stay in Google's results are never returned again.
• Don't run two alert runs with the same Tracking ID at the same time; the later one wins when saving.

Getting more than 100 articles per query

Each Google News feed returns at most about 100 articles (we saw 100 to 104), ranked by relevance. With Split long time ranges by day on, a search with a 24-hour, 7-day, 30-day or custom date range is fetched as one feed per day, newest day first, and the Actor stops as soon as Max articles per query (or your remaining spending limit) is reached. In our test, Tesla over 7 days returned 711 unique articles from 8 requests in 5 seconds. Topic, top-story and local feeds can't be split (they have no date operators).

Publisher URLs (optional, best effort)

Google News RSS links point to news.google.com/rss/articles/..., which redirects to the publisher in a browser. With Resolve publisher URLs on, the Actor also tries to decode them to the publisher's own URL.

• It uses an undocumented Google News web endpoint, the same one the Google News site uses. Google can change or block it at any time, so this feature may stop working without notice.
• It fails soft: if decoding fails, the article is still returned with its googleNewsUrl and publisherUrl: null, and the run never fails because of it. After two failed batches in a row the Actor stops trying for the rest of the run and says so in the SUMMARY record.
• In our cloud tests on 19 and 20 Oct 2026, every link was decoded (200 of 200, from Apify servers, directly and through the datacenter proxy).
• It costs about 1 extra request per article (one page of about 125 KB, plus one small request per 10 articles), so runs take a little longer: roughly 0.5 to 3 seconds per 10 articles (measured: 120 articles decoded in 6 seconds).
• It's off by default. If you only need to open articles, googleNewsUrl works in any browser.

Countries and languages

71 countries and 40 languages, verified against Google News (each combination is a Google News "edition"). Examples: United States (English, Spanish), United Kingdom, India (English, Hindi, Bengali, Marathi, Tamil, Telugu, Malayalam, Gujarati), Canada (English, French), Germany, France, Spain, Italy, Brazil, Mexico, Japan, South Korea, China, Taiwan, Indonesia, Turkey, Saudi Arabia, Israel, Nigeria, South Africa and many more. Leave Language on Auto to use the country's main language.

If you pick a language a country has no edition for (for example Germany + English), the Actor uses the country's main language and notes it in the run summary. (Google itself would silently switch you to a different country.)

Pricing

Pay per event, no subscription and no extra usage fees:

Event	Price
article: one article in the dataset	$0.0015 ($1.50 per 1,000)
Actor start (standard Apify event, once per run)	$0.0005

Examples:

• 10 queries, past 24 hours, 100 articles each: 1,000 articles = $1.50.
• Hourly alert run for 5 brand queries that finds 3 new articles per hour on average: 72 articles plus 24 runs a day = about $0.12 per day. A run that finds nothing new costs only the start event ($0.0005), so an hourly schedule that never finds anything costs about $0.36 a month.
• The prefilled example (2 queries, 10 articles each): about $0.03 ($0.0305).

Already-seen articles, skipped duplicates, error rows and the SUMMARY record are free. Set Maximum cost per run in the run options (or maxTotalChargeUsd in the API) to cap spending: the Actor stops cleanly at the cap, never fetches feeds it can't charge for, and everything returned so far stays in the dataset.

Use it from the API

Run synchronously and get the articles in one call (keep your Apify API token in the APIFY_TOKEN environment variable):

curl -X POST "https://api.apify.com/v2/acts/fairfetch~google-news-monitor/run-sync-get-dataset-items?maxTotalChargeUsd=0.30" \
  -H "Authorization: Bearer $APIFY_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"queries": ["Nvidia earnings"], "timeRange": "1d", "maxArticlesPerQuery": 20}'

The items include free error rows when a query fails; filter on status == "ok" if you only want articles. Each run also writes a SUMMARY record with per-query counts, the edition used, stop reasons and any errors.

Performance and limits

Measured on the Apify platform (256 MB, 19 and 20 Oct 2026):

• The prefill (2 queries x 10 articles): 20 articles, 3.9 seconds in total.
• 6 queries x 20 articles with publisher URLs: 120 articles, all 120 publisher URLs decoded, 6 seconds of work, 15 MB downloaded.
• 3 queries x 5 articles with publisher URLs: about 1.5 seconds of work with direct requests, 6 to 12 seconds through Apify's datacenter proxy.

Limits to know:

• About 100 articles per feed (use Split long time ranges by day for more). Top stories return about 35 to 40 articles, topic and local feeds about 60 to 90 (some older than 24 hours, which the time filter removes).
• Google ranks search feeds by relevance; Order: Newest first sorts them by date before applying Max articles per query.
• Requests are polite: at most 3 at a time by default, retries with exponential backoff (longer when Google signals rate limiting), and timeouts.
• If Google blocks a run, the Actor switches to Apify's datacenter proxy with a fresh IP for each retry. If Google keeps blocking (10 blocked answers in a row), it stops sending requests, reports the remaining feeds as blocked and ends the run instead of waiting until the timeout.
• Run timeout: the platform needs up to about 7 seconds to finish a run after the Actor exits, so the Actor stops reading new feeds 16 to 70 seconds before the run's timeout (depending on the timeout) and stops pushing rows 12 to 20 seconds before it. Rows that were read but not pushed are not charged, and the SUMMARY is always written. Tested: 20-, 30- and 45-second timeouts with large feeds end SUCCEEDED. The Actor needs a timeout of at least 20 seconds: with less, it stops at once with "Run timeout too short" and a free error row. Use 30 seconds or more if you set a timeout.

FAQ

Does it return the full article text? No. Google News RSS feeds contain the headline, publisher, publish time and a link, not the article body. Use publisherUrl (or open googleNewsUrl) to read the article on the publisher's site, subject to the publisher's terms and paywalls.

Why is the snippet so short? Google no longer includes article summaries in its RSS feeds. For search feeds the description is just the headline and publisher; for top stories and topic sections it lists related headlines (also available as relatedArticles). The Actor returns that text as clean plain text and doesn't invent summaries.

Does it include images, authors or article categories? No, the feeds don't contain them.

Why did I get fewer articles than "Max articles per query"? Google returned fewer for that query and time range, the time filter removed older items, or the articles were duplicates of another query or already seen in alert mode. The SUMMARY record shows the count for each reason.

How fresh is the data? Feeds are read live on every run. New articles usually appear in Google News within minutes of publication.

Do I need proxies or a Google account? No. The feeds are public and the Actor reads them without signing in. Requests go out directly; if Google blocks a run on the Apify platform, the Actor switches to Apify's datacenter proxy by itself. You can also choose Apify Proxy (datacenter) or your own proxy URLs. Residential and Google SERP proxy groups are not supported: a run that selects them stops at once with "Invalid input" and no articles are charged (only the standard Actor start event).

Can I monitor many queries? Yes, there is no hard limit. Feeds are fetched 3 at a time by default and each takes well under a second in our tests, so even long query lists finish quickly (publisher URL decoding is the slow part).

What happens if one query fails? The others still run. The failed one gets a free error row with the reason, and it's listed in the SUMMARY record and the status message. A run where every entry was wrong (for example only unknown places) still ends as SUCCEEDED, with the error rows; a run ends as FAILED only when nothing could be returned because of a problem on our side, such as Google blocking every request.

What's new in 1.1 (October 2026)

• Free error rows: a query, topic, place or URL that can't be read now adds a row with status: "error", error and errorKind (before: only in the SUMMARY). Every row has a status field; filter on status == "ok" if you only want articles. New dataset view: Errors.
• Input mistakes end SUCCEEDED: a run whose entries are all wrong (for example only unknown places) or whose settings are invalid now ends SUCCEEDED with error rows and an "Invalid input" or "No articles returned" status, instead of FAILED. One bad URL no longer stops the other queries, and a google.com News-tab URL is turned into a search query.
• Tracking ID: it has no default value any more. Without one, each saved task (or each set of queries plus country and language) gets its own alert memory; monitors that ran before 1.1 without an ID keep what they already received.
• Time limit, spending limit, blocks and migrations: the Actor stops cleanly before the run's timeout, never fetches what your spending limit can't pay for, switches to a datacenter proxy if Google blocks it, and never repeats or re-charges articles after a platform migration.
• Proxies: only datacenter proxies (Apify's or your own); residential and Google SERP groups are refused.
• Price: articles still cost $0.0015 each; the per-run start fee is now $0.0005 (was $0.00005), so runs that find nothing cover their own platform cost.

Legal

• Unofficial. This Actor is not affiliated with, endorsed by or sponsored by Google. Google News is a trademark of Google LLC; all other names are the property of their owners.
• What it reads. Only public Google News RSS feed URLs, without signing in, plus (only when you turn on Resolve publisher URLs) Google News article pages and an undocumented Google News endpoint to find publisher links. It doesn't bypass paywalls or access controls and doesn't collect personal data beyond what appears in public headlines (such as names of public figures).
• Google's terms. Every Google News feed carries a copyright notice saying the feed is made available for rendering results in a personal feed reader for personal, non-commercial use, and that other uses are prohibited. Google News' robots.txt does not allow automated access to the RSS feed paths. The headlines and articles belong to their publishers.
• Your responsibility. You decide what you do with the data. Check that your use complies with Google's terms, the publishers' rights and the laws that apply to you (such as copyright and data-protection law), and get legal advice before any commercial use or redistribution.